Factorization machines (FMs) are a supervised learning approach that can use second-order feature combinations even when the data is very high-dimensional. Unfortunately, despite increasing interest in FMs, there exists to date no efficient training algorithm for higher-order FMs (HOFMs). In this paper, we present the first generic yet efficient algorithms for training arbitrary-order HOFMs. We...

The majority of real-world networks are dynamic and extremely large (e.g., Internet Traffic, Twitter, Facebook, ...). To understand the structural behavior of nodes in these large dynamic networks, it may be necessary to model the dynamics of behavioral roles representing the main connectivity patterns over time. In this paper, we propose a dynamic behavioral mixed-membership model (DBMM) that ...

In axiomatic way, in fuzzy setting, we introduce the measure of crispness and the measure of conditional clearness. For these measures we propose a functional form. Then, for the last measure, we give some aggregation operators. Keywords— Aggregation operators, functional equations, fuzzy sets, measures of fuzziness, crispness measures, conditional crispness.

We present a method of producing normal subgroups of the free group on r generators not containing a given word. Special attention is given to the normal subgroups of the free group on two generators. In considering the exponent sum of the generators in a given word, we produce such subgroups and clearly state for what class of words these subgroups are not easily obtainable. We also provide a ...

The determination of the mean survival time t ~i.e., inverse reaction rate! associated with diffusion-controlled reactions among static traps is a problem of long-standing interest, dating back to the classical work of Smoluchkowski. For the broad class of model particulateand digitized-based models considered here, we find a universal curve for the mean survival time t for a wide range of poro...

For at least partially ordered three-way tables, it is well known how to arithmetically decompose Pearson’s X2 P statistic into informative components that enable a close scrutiny of the data. Similarly well-known are smooth models for two-way tables from which score tests for homogeneity and independence can be derived. From these models, both the components of Pearson’s X2 P and information a...

We analyze the general structure of soft scalar masses emerging in superstring models involving anomalous U(1) symmetries, with the aim of characterizing more systematically the circumstances under which they can happen to be flavor universal. We consider both heterotic orbifold and intersecting brane models, possibly with several anomalous and non-anomalous spontaneously broken U(1) symmetries...

Previously, the weight parameter of the finite mixture of regression models has been assumed to be invariant of the characteristics of the observations under study. Recently, it has been shown that the weight parameter of the finite mixture of negative binomial (NB) models can be dependent upon the attributes of the sites. Since the selection of the functional form for weight parameter has a si...

The evolution with energy scale of the partonic logarithmic scaled energy spectra is investigated in the framework of the modified leading logarithmic approximation (MLLA). The behaviour of the higher order moments is compared to a number of analytic predictions and e + e − data.